PCCP

View Article Online

PAPER View Journal | View Issue
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

From information theory to quantitative

description of steric eects
Cite this: Phys. Chem. Chem. Phys.,
2016, 18, 17917
Mojtaba Alipour* and Zahra Safari

Received 25th April 2016,
Accepted 6th June 2016
DOI: 10.1039/c6cp02750d
www.rsc.org/pccp
Immense eorts have been made in the literature to apply the information theory descriptors for
investigating the electronic structure theory of various systems. In the present study, the information
theoretic quantities, such as Fisher information, Shannon entropy, Onicescu information energy, and
GhoshBerkowitzParr entropy, have been used to present a quantitative description for one of the
most widely used concepts in chemistry, namely the steric effects. Taking the experimental steric scales
for the different compounds as benchmark sets, there are reasonable linear relationships between
the experimental scales of the steric effects and theoretical values of steric energies calculated from
information theory functionals. Perusing the results obtained from the information theoretic quantities
with the two representations of electron density and shape function, the Shannon entropy has the
best performance for the purpose. On the one hand, the usefulness of considering the contributions of
functional groups steric energies and geometries, and on the other hand, dissecting the effects of both
global and local information measures simultaneously have also been explored. Furthermore, the utility
of the information functionals for the description of steric effects in several chemical transformations,
such as electrophilic and nucleophilic reactions and hostguest chemistry, has been analyzed. The functionals
of information theory correlate remarkably with the stability of systems and experimental scales. Overall, these
findings show that the information theoretic quantities can be introduced as quantitative measures of steric
effects and provide further evidences of the quality of information theory toward helping theoreticians
and experimentalists to interpret different problems in real systems.

1 Introduction Aiming to develop quantitative measures in this context, a

Steric eects, as one of the essential and most widely used
concepts in chemistry, have long been known to play important
roles in understanding the behavior of dierent phenomena.
The steric eect is basically linked to this fact that atoms
and molecules occupy certain space volumes. The hindrance
induced when they are brought together change the energy,
shape, reactivity, and other properties of the interacting particles.
These changes are associated with the concept of steric eect.1
Among the abundant processes in which the steric eects have
imperative roles, the rate and energy of chemical reactions, the
structure, dynamics, and function of enzymes, and the inter-
action of a drug molecule with a receptor can be noted to
mention a few cases. Notwithstanding the plethora of qualitative
considerations on the steric eects in the literature, like many
other concepts in science, there is no precise quantification
description of this substantial eect. In fact, there is no
physically observable variable associated with the steric eect
and therefore quantum mechanically speaking, it is a noumenon.27
Department of Chemistry, College of Sciences, Shiraz University, Shiraz, Iran.
E-mail: malipour@shirazu.ac.ir
density functional theory (DFT)810 description of the steric
effects was proposed by Liu.11 This approach is based on the a
new energy partition scheme, in which the total electronic
energy is decomposed into contributions from three independent
parts: steric, electrostatic, and fermionic quantum effects. It has
pragmatically been proven that the steric energy, ES[ r], as a
functional of electron density can be expressed quantitatively
as follows:
1 jrrrj2
ES r TW r dr (1)
8 rr
where TW[r] is the Weizsacker kinetic energy.12 The physical
nature of this new quantification of the steric effect is based on
the consideration of a new reference state in which electrons of
atoms and molecules are assumed to behave like bosons. If the
density of the hypothetical boson state is the same as that of the
fermionic state, r(r), the total wave function of the hypothetical
p
state will be rr=N (N is the number of electrons).11 The total
kinetic energy of the hypothetical state, from which Weisskopfs
kinetic energy pressure for the steric effect is calculated,13 is
simply eqn (1). Following the original study on the quantitative

This journal is the Owner Societies 2016 Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17917
 View Article Online

Paper PCCP

viewpoint of the steric effects, this scheme has been employed
in a number of topics in recent years.1433 We have also
proposed another version of quantitative description of steric
effects in momentum space and tested the resulting expressions
for some phenomena, in which the steric effects can be
important.23,25
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
measure of the sharpness or localization of the electron density.
As a matter of fact, considering the discrete forms of the Shannon
entropy and Fisher information instead of their definitions in
continuous forms, using the mathematical arguments on the
behaviour of their expressions in terms of the probability
distributions, and analyzing the continuity and discontinuity

From a dierent perspective, the definition of steric energy of their graphical curves the global and local natures of the
in eqn (1) is related to the functionals of information theory.3438 Shannon entropy and Fisher information, respectively, can be
Information theory is a branch of applied mathematics, electrical justified.58 Moreover, for Fisher information, eqn (2), an equivalent
engineering, and computer science that deals with the quantifi- expression in terms of the Laplacian of the electron density has
cation of information, which is often a probability distribution also been proposed59
function. In recent years, considerable effort has been invested to
0
understand and describe the physical and chemical processes in IF r  r2 rr ln rrdr (6)
the framework of information theory, observe for instance
ref. 3957. On the other hand, considering this point that the On the other hand, the information theoretic quantities defined
ground state electron density determines everything, various in eqn (2)(6) include the electron density as the probability
information theoretic functionals should also be related. Putting distribution function. However, there is another distribution
these perceptions together, we anticipate that the quantitative function, namely, the shape function s(r),60 upon which the
description of the steric effect can be rationalized based on the information functionals can be defined. s(r) and r(r) are related
information theoretic quantities. Accordingly, scrutinizing the through the following relation:
related functionals and numerical analyses to confirm this issue r(r) = Ns(r) (7)
and consequently comparing the results of steric energies
obtained from different information theoretic functionals for with the normalization condition srdr 1. For some con-
various chemical transformations, in which steric effects come siderations about the shape function in connection to the
into play, are the subjects of the present contribution. electron density we refer to a few studies in this context.6164
Accordingly, the information theoretic quantities can be
rewritten based on the shape function as follows:
2 Theoretical considerations
jrsrj2
The central functionals of information theory of concern are the IF s dr (8)
sr
Fisher information IF[r],35 Shannon entropy SS[r],36 Onicescu infor-

mation energy EO[r],37 and GhoshBerkowitzParr (GBP) entropy
S S s  sr ln srdr (9)
SGBP[r].38 These functionals are defined as follows:

jrrrj2
IF r dr (2) EO s sr2 dr (10)
rr
 
S S r  rr ln rrdr (3) 3 tr;s
S GBP s ksr c ln dr (11)
2 tTF r;s

EO r rr2 dr
0
(4) IF s  r2 sr ln srdr (12)

  Imposing the relation between the electron density and shape

3 tr;r
S GBP r krr c ln dr (5) function, eqn (7), it can be shown easily that the information
2 tTF r;r
theoretic quantities in the two representations s(r) and r(r) are
with r(r) as the ground state electron density of an N-electron related through the following equations:

system, which satisfies the normalization condition rrdr N. IF r
5 4pck IF s (13)
In eqn (5), k is the Boltzmann constant, c ln , t(r;r) is the N
3 3
kinetic energy density, which is related to the total kinetic energy TS S S r
S S s ln N (14)
via tr;r dr TS , and tTF(r;r) is the ThomasFermi kinetic energy N
3  2 2=3
density given by tTF(r;r) = ck[r(r)]5/3 with ck 3p .
10 EO r
EO s (15)
Several points about the functionals of information theory N2
are of concern. From the viewpoint of information theory, the
0
Shannon entropy is a global measure of the spatial delocalization 0 IF r
IF s (16)
of the electron density, whereas Fisher information is a local N

17918 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 This journal is the Owner Societies 2016
 View Article Online

PCCP Paper

On the GBP functional, it should be noted that because the
scaling properties for the kinetic energy density t(r;r) are
unknown, it is unclear if there are any explicit relationships
between SGBP[r] and SGBP[s]. On the other hand, it has numerically
and analytically been proven that the two GBP functionals are
connected as follows:65,66
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
has been used for the present calculations. However, following
previous suggestions and to have straightforward comparisons
between these results with earlier published data in this
context, the appropriate basis sets were employed for geometry
optimizations and subsequent calculations.17,83 In particular,
for CH3X compounds and esters, for which the experimental

scales of the steric effects were available, the 6-311+G(2d,2p)

S GBP r
S GBP s (17)
N
Comparing the definition of steric energy in eqn (1) and the
information theoretic functionals, we find that the steric eect
is intrinsically related to the information theory. In more detail,
eqn (1) without a multiplying factor of 1/8 is the definition
of Fisher information, eqn (2), and therefore the steric energy
1 were implemented in the multifunctional wave function analyzer
can be rewritten as ES r IF r. On the other hand, it has
8 program developed by Lu and Chen,85 which its applicability and
been proven earlier that there are relationships among various
functionals of information theory.59,67,68 Therefore, it is expected
that the quantitative description of the steric eect can also be
proposed based on other information functionals and not only
based on Fisher information. In fact, it would be interesting to
observe whether or not there are other definitions for steric
energy, other than eqn (1), in the two representations of s(r)
and r(r) with superior performance in applicability and account-
ability for quantitative analysis of steric eects. We will provide
an armative ratification for this issue in the present study.
3 Computational details
As systems under study, we first consider compounds with the
experimental scale of steric eects introduced by Taft.6972
More specifically, we have chosen a set of first row and second
row CH3X compounds with various electron donating and
electron withdrawing groups X. In addition, a number of esters
containing the cyclic functional groups prone to steric analysis
have also been selected. After direct comparison between
experimental and theoretical scales of steric energies, the
accountability of information theoretic quantities in unveiling
the steric effects is explored employing several processes, in
which the steric effects play a leading role. To this end, a few
benchmark sets have been considered. First, as illustrative
examples of electrophilic and nucleophilic aromatic substitution
reactions, we considered the two sets of molecules containing
various aromatic rings and substituent groups with experimental
electrophilicity and nucleophilicity scales introduced by Mayr.7376
and aug-cc-pVTZ basis sets have been used, respectively. On the
other hand, for consistency, all the subsequent calculations
about exploring the role of steric effects in larger systems of the
benchmarked sets were performed using the 6-311++G(d,p)
basis. The corresponding wave functions for all the systems
were derived using the Gaussian 09 suite of codes.84 Moreover,
all the runs to compute the information theoretic functionals
reliability has been extensively advocated before. Atomic units
were used throughout.
4 Results and discussion
We first evaluated the possibility of establishing links among
the theoretical values of the steric energies obtained from the
information functionals and experimental scales. Table 1 lists
the considered first row and second row CH3X molecules and
esters containing cycloalkyl groups along with their experimental
steric values. Fig. 1 and 2 show the corresponding linear
correlations between the theoretical and experimental scales of
steric effects for the CH3X compounds and esters, respectively.
At first glance, it can be observed that there are linear
correlations between the two scales for the both sets of CH3X
compounds and esters. Furthermore, the best correlation is
achieved using the Shannon entropy, followed by GBP entropy,
Fisher information, and Onicescu information energy. Upon
a closer inspection in the figures, we find that the slope is
negative in all correlations, indicating that larger values of
information theoretic quantities lead to more negative experi-
mental scales. It should be noted that there are discrepancies
between the experiment and theory, which may be attributed to
Table 1 Tafts steric values for CH3X compounds and CH3OCOHR
esters considered in this work
X ETaft
S R ETaft
S

Furthermore, the performance of the information theoretic H 0.00 Cyclobutyl 0.06

functionals was further assessed for the chemistry of hostguest F 0.46 Cyclopentyl 0.51
complexes involving the interaction between large host compounds CN 0.51 Cyclohexyl (eq) 0.79
OH 0.55 Cyclohexyl (ax) 0.79
with small guest molecules.77,78 For this purpose, hostguest OMe 0.55 Cycloheptyl 1.10
complexes containing polycyclic aromatic hydrocarbons as NH2 0.61 C(CH3)(cyclohexyl) 2.03
guests and the ExBox compounds as host systems79,80 are Me 1.24 C(CH3)2(cylohexyl) (eq) 2.49
Et 1.31 C(CH3)2(cylohexyl) (ax) 2.49
considered where their stability and other related aspects have Pr 1.60
been studied from the viewpoint of steric effects analysis. It has i-Pr 1.71
previously been pointed out that the quantitative steric analyses i-Bu 2.17
s-Bu 2.37
do not depend on the choice of level of theory. Therefore, the t-Bu 2.78
widely used exchangecorrelation DFT functional, B3LYP,81,82 Neopentyl 2.98

This journal is the Owner Societies 2016 Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17919
 View Article Online

Paper PCCP
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

Fig. 1 Correlations between the theoretical steric energies computed from information theoretic quantities and Tafts steric values for the CH3X
compounds under study. The fitting expressions and regression coefficients are also given.

Fig. 2 Correlations between the theoretical steric energies computed from the information theoretic quantities and Tafts steric values for the esters
under study. The fitting expressions and regression coecients are also given.

17920 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 This journal is the Owner Societies 2016
 View Article Online

PCCP Paper
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

Fig. 3 Correlations between the theoretical steric energies computed from the shape function-based Shannon entropy and Tafts steric values for (a)
CH3X compounds and (b) esters.

several factors. Among others, the different conditions of Correlation issues aside, the dierent behaviours of the
experimental measurements and theoretical predictions can information theoretic functionals in the prediction of steric
be mentioned. For example, the experimental scales are from eects are aected from their integrands. Fig. 4 displays the
the acid-catalyzed hydrolysis of esters, while the theoretical local distributions of the Fisher information (eqn (2)), Shannon
results have been calculated in the gas-phase. On the other entropy, Onicescu information energy, and GBP entropy for the
hand, similar analyses have also been carried out using the CH3X molecule with X = isopropyl (isobutane) as illustrative
information theoretic functionals with shape function. It is examples. We can observe from the figure that the local
found that the correlations obtained using the shape function- behaviours of the Fisher information and Onicescu information
based information functionals are not as strong as those energy (and somewhat GBP entropy) are similar and sub-
provided from the conventional informational theoretic quantities. stantially different from that provided by Shannon entropy.
Nevertheless, only the Shannon entropy provides still the best Moreover, the spatial delocalization of the electron density in
performance among the others for the considered sets (see Shannon entropy and the sharpness or localization of the
Fig. 3 for the corresponding correlations derived from the shape electron density in Fisher information are evident from their
function-based Shannon entropy for the considered CH3X scattered and concentrated distributions in Fig. 4, which is
compounds and esters). consistent with earlier findings in this context.28 It may there-
We also examined the correlation among information theoretic fore be concluded that for investigations on the global steric
quantities based on Pearsons correlation matrix analysis. The effects, the Shannon entropy is a good choice, whereas the
corresponding Pearsons correlation matrix for all the considered Fisher information is a better option for analyzing the local
information functionals has been provided in Table 2. In (atomic and functional group contributions) analysis of the
agreement with what was shown analytically on the relation- steric effects. This is in agreement with what was previously
ships among informational functionals,59,67 it is quite obvious shown by Torrent-Sucarrat et al.17
from the table that in most cases, the correlation among the Interestingly, the role of the integrands is more pronounced
tested information measures is high, especially among infor- when comparing the two equivalent forms of the Fisher infor-
mation theoretic functionals for the esters under study, which mation, eqn (2) and (6). Shown in Fig. 5 are the local distributions
in turn shows the consistency between the results of functionals of the two forms of Fisher information for isobutane. In agreement
of information theory. with what has earlier been pointed our for atoms,59 we find that
although the two forms of the Fisher information are identical in
value, there are significant dierences in the local behaviours of
Table 2 Pearsons correlation matrix on the considered information their integrands and consequently in their predictions of steric
theoretic quantities for CH3X compounds and esters under study eects.
Information theoretic quantities IF[r] SS[r] EO[r] SGBP[r] In another tack, the contribution of a functional group and
the relevance of the geometry, on the one hand, and the
CH3X compounds
IF[r] 1.000 applicability of both information theoretic quantities simulta-
SS[r] 0.868 1.000 neously, on the other hand, have also been taken into account.
EO[r] 0.944 0.665 1.000 To this end, we performed various multilinear regression
SGBP[r] 0.960 0.972 0.818 1.000
analyses. As global and local information theoretic functionals,
Esters the Shannon entropy and Fisher information are considered for
IF[r] 1.000 multilinear regressions, respectively. In the first case, bilinear
SS[r] 0.999 1.000
EO[r] 1.000 0.999 1.000 regression analysis has been performed using the two calculated
SGBP[r] 1.000 0.999 0.999 1.000 parameters, CH3 steric energies and CX bond distance, for the

This journal is the Owner Societies 2016 Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17921
 View Article Online

Paper PCCP
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

Fig. 4 Local distributions of the Fisher information (a), Shannon entropy (b), Onicescu information energy (c), and GBP entropy (d) for the isobutane molecule.

Fig. 5 Comparison of the local distributions of the (a) Fisher information defined in eqn (2) and (b) Fisher information defined in eqn (6) for the isobutane molecule.

CH3X compounds wherein following previous suggestions only ETaft

S = 3.991SS[r]CH3  62.406dCH3X + 40.734 (18)
the systems with X = C are considered.17 The same analyses have
also been performed using the parameters CH3OCOH steric ETaft
S = 2.462IF[r]CH3  25.328dCH3X  595.087 (19)
energies and CH3OCOHR bond distance for the esters under ETaft = 2.580SS[r]CH3OCOH  27.541dCH3OCOHR + 97.075 (20)
S
study. As a result, the following bilinear expressions were
derived: ETaft
S = 0.472IF[r]CH3OCOH  44.299dCH3OCOHR  596.295 (21)

17922 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 This journal is the Owner Societies 2016

PCCP
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
Fig. 6 Correlations between the theoretical steric energy computed from eqn (18)(21) and Tafts steric values. The fitting expressions and regression
coefficients are also given.
To validate the applicability of these equations, the correlation
plots between the Taft experimental scale and the fitted theoretical
steric energies obtained from eqn (18)(21), denominated as
ETaft Taft Taft Taft
S (18), ES (19), ES (20), and ES (21) are presented in Fig. 6.
Compared to the discussed linear correlations, we can
observe from Fig. 6 that in most cases, the improved correlations
are obtained using the bilinear equations. Thus, the account-
ability of information theoretic functionals is confirmed not only
for the global steric eect analyses, but also for investigating the
eects of atomic and functional group contributions in steric
energies. Moreover, consistent with the previously published
results in this respect,17 it seems that considering the contributions
from functional groups and molecular geometries are essential in
quantitative steric analyses. For the second case, the applicability
of both information theoretic functionals simultaneously,
Fig. 7 Correlations between the theoretical steric energy computed from eqn (22) and (23) and Tafts steric values. The fitting expressions and
regression coecients are also given.
This journal is the Owner Societies 2016
View Article Online
Paper
other bilinear regression analyses have also been carried out
using the Shannon entropy (global measure) and Fisher infor-
mation (local measure) for both sets. The resulting expressions
on the CH3X compounds and esters are respectively as follows:
ETaft
S = 0.047SS[r] + 0.0001IF[r] + 0.597 (22)
ETaft
S = 0.958SS[r] + 0.037IF[r]  32.263 (23)
The fitted steric energies based on eqn (22) and (23), denominated
as ETaft Taft
S (22) and ES (23), respectively, against experimental scales
of steric eects are plotted in Fig. 7.
The obtained results show that mixing the global and local
information theoretic quantities provides good correlations.
In more detail, the relationships are improved when going
from a linear correlation to bilinear expression containing the
Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17923

Paper
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
Fig. 8 Structures of the compounds considered for electrophilicity analysis. Colour conventions: gray (carbon), light gray (hydrogen), red (oxygen), blue
(nitrogen), and turquoise (fluorine).
Shannon entropy and Fisher information, highlighting the use
of both global and local information functionals for that
purpose. These findings prompt us to perform comprehensive
analyses for all the possible combinations among the infor-
mation theoretic quantities. All the bilinear regressions led to
good correlations between the two scales. In addition, we found
that for cases in which information functionals alone do not
provide satisfactory correlations, imposing the second variable
17924 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929
View Article Online
PCCP
in bilinear regressions ameliorates the results. On the whole,
our results unveil that in addition to Fisher information other
information theoretic quantities based on electron density can
also be considered as a tool to quantify the steric eects. Thus,
for the processes in which the steric eects are important, there
should be a connection between their indices and information
theoretic quantities. We now deal with such relationships using
several benchmark sets. As a case study, we investigated the
This journal is the Owner Societies 2016
 View Article Online

PCCP Paper

quantities and consequently of the strain, the larger the electro-

philicity scale and the better electron-accepting capability at
the reaction center, as anticipated from the chemical intuition.
On the other hand, perusing the relationships among the
information functionals and nucleophilicity scale (Fig. 11), we
find that the information theoretic quantities are positively
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

proportional to the nucleophilicity data. In this case, however,

other information functionals provide a better correlation compared
to Shannon entropy. In a related study, Liu et al.86 recently proposed
a unified framework, called the information conservation principle,
to determine the electrophilicity, nucleophilicity, and regioselectivity
scales. In agreement with our findings, it has been shown that there
are also strong correlations among another information theoretic
quantity, i.e., information gain (information divergence, relative
entropy, KullbackLeibler divergence),87 Hirshfeld charges, and
experimental scales of electrophilicity and nucleophilicity.
As a last test case, we analyzed the steric eects in the host
guest complexes. First, we focused on some recently proposed
host molecules, ExBox compounds, for the purpose.88,89 Fig. 12
shows the two types of ExBox molecules, (a) and (b), wherein
Fig. 9 Structures of the compounds considered for nucleophilicity analysis.
Colour conventions: gray (carbon), light gray (hydrogen), red (oxygen), and the case (b) has been proposed as a flatter molecule with
blue (nitrogen). respect to the case (a) and provides a less congested interior
to contain the guest molecules.79,80
From the viewpoint of steric analysis, it is anticipated that
electrophilic and nucleophilic reactions as an important the steric energy in compound (b) is smaller than that of
chemical transformation in chemistry, emphasizing the steric compound (a). The computed steric energies using all the
eects. The considered sets of compounds for electrophilicity information functionals considered for both compounds are
and nucleophilicity analyses are shown in Fig. 8 and 9, respectively, gathered in Table 4. Zooming in the obtained data, we find that
and the corresponding experimental scales are collected in Table 3. although all the information theoretic functionals do not
In addition, we plot in Fig. 10 and 11 the relationships between provide the correct order for the steric energies, the Shannon
information theoretic quantities and experimental data for the entropy shows the expected ordering wherein the Shannon
two sets. entropy for compound (b) is less than that provided for
Fig. 10 shows that there exist linear correlations between the compound (a). We now examined the hostguest complexes
electrophilicity scale and information theoretic quantities formed from these host compounds and aromatic hydro-
wherein the obtained correlation from Shannon entropy is better carbons as a guest. It is noted here that because the aromatic
than the others. More importantly, the electrophilicity scale is hydrocarbons can be toxic and carcinogenic, their isolation is
negatively proportional to the amount of the informational important for environmental reasons. As suggested in previous
functionals. The smaller the amount of the information theoretic experimental and theoretical investigations,79,80,88,89 the ExBox
host molecules with an excellent environment for capturing the
aromatic hydrocarbons can be used for such isolation process.
Table 3 Experimental electrophilicity and nucleophilicity scales for mole- The guest molecules considered to form hostgust complexes
cules considered in this work are benzene, naphthalene, and anthracene. Shown in Fig. 12
are the structures of the resulting hostguest complexes under
Molecule Electrophilicity Molecule Nucleophilicity
study79,80 and the corresponding steric energies obtained from
I 5.9 I 4.5 information functionals are reported in Table 4.
II 5.6 II 1.2
III 4.6 III 0.13 A glance to this data is sucient to conclude that when
IV 3.6 IV 2.5 going from benzene to naphthalene to anthracene, the computed
V 2.1 V 7.3 steric energies increase for both host(a)guest and host(b)guest
VI 1.5 VI 6.3
VII 0.0 VII 0.02 complexes, as expected qualitatively. On the other hand, com-
VIII 7.0 VIII 3.4 paring the two complexes for each guest, we find again that
IX 7.7 among all the tested information theoretic quantities, only
X 8.8
XI 10.0 the Shannon entropy provides the correct order, wherein
XII 11.9 the computed steric energies from the Shannon entropy for
XIII 15.0 host(b)guest complex with a flatter host molecule are smaller
XIV 15.8
XV 16.1 than that provided for the host(a)guest complex. However,
XVI 17.3 although all the information theoretic quantities do not describe

This journal is the Owner Societies 2016 Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17925
 View Article Online

Paper PCCP
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

Fig. 10 Relationships between the information theoretic quantities and experimental electrophilicity scales. The fitting expressions and regression
coecients are also given.

Fig. 11 Relationships between the information theoretic quantities and experimental nucleophilicity scales. The fitting expressions and regression
coecients are also given.

the behavior of the steric effects quantitatively, the quantities As an example, from the definition of steric energy, the steric
derived from their definitions may reveal the appropriate trends. force can also be introduced,19

17926 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 This journal is the Owner Societies 2016
 View Article Online

PCCP Paper

system due to the stress factor and its integral over the atomic
basin will yield the whole steric force acting on a bounded
atom. As an application of the steric force, we refer to a detailed
investigation by Tsirelson et al.19 on the diamond and boron
nitride crystals. The formalism of steric force using the definition
of steric energy based on Fisher information has been employed
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.

in our calculations as a representative example. Using eqn (1) for

obtaining the steric potential us, the expression of the steric force
is obtained as follows:
" #
1 jrrrj2 1 r2 rr
fs r  (25)
8 rr2 4 rr

The values of the steric force obtained from eqn (25) for all the
considered hostguest complexes are reported in Table 4 (see the
data in parentheses of the second column). As can be observed
from these results, even though the steric energies obtained
based on Fisher information do not confirm the order of strain
for the two complexes with each guest, the values of steric force
provide a good description in this case. Indeed, for each of the
guest molecules, when going from host(a)guest complex to
host(b)guest complex, the values of the steric force decrease,
as expected, whereas the values of the steric energies based on
Fisher information show an increasing pattern. Therefore, it can
be concluded that the steric force can also be introduced as a
descriptor to quantify the steric effects in hostguest complexes.

5 Conclusion
During this study, the accountability of informational theoretic
functionals, such as Fisher information, Shannon entropy,
Onicescu information energy, and GhoshBerkowitzParr entropy,
has been benchmarked for quantitative analysis of the global
Fig. 12 Structures of ExBox host molecules (a and b) and their complexes and local steric effects. By applying these information theory
with benzene, naphthalene, and anthracene. Colour conventions: gray functionals to several benchmark sets, interesting templates
(carbon), light gray (hydrogen), and blue (nitrogen).
were disclosed. Considering the experimental steric values for
different compounds, such as substituted methanes and esters,
Table 4 Computed values of steric energies using all the information we observed reasonable linear relationships between the experi-
theoretic quantities for the host molecules and hostguest complexes. mental scale of the steric effects and the theoretical values
The values in parentheses refer to the steric force obtained from eqn (25), computed from information theoretic quantities with the
see text
two representations of electron density and shape function.
IF[r] SS[r] EO[r] SGBP[r] Dissecting the obtained correlations and the behaviors of
the integrands of the functionals, we found that the best
Host molecules
(a) 1655.5 331.0 1727.2 2383.4 correlation is derived from Shannon entropy and consequently
(b) 1747.4 299.3 1893.7 2386.7 this measure of information theory can be introduced as a
Hostguest complexes new tool to quantify steric analyses. It has also been revealed
(a)benzene 1844.0 (10138.5) 375.1 1917.0 2667.4
(b)benzene 1933.5 (9966.4) 343.3 2083.6 2670.7 that for the detailed analyses in this context, considering the
(a)naphthalene 1964.4 (10638.1) 398.2 2043.4 2842.9 contributions of functional groups steric energies and geo-
(b)naphthalene 2056.4 (10489.8) 366.4 2210.0 2846.2 metries and the effects of both global and local information
(a)anthracene 2087.2 (11142.1) 420.9 2169.8 3018.2
(b)anthracene 2179.0 (10976.0) 389.0 2336.3 3021.3 functionals simultaneously within multilinear regression analyses
is essential. Confirming the applicability of information theoretic
quantities as measures to quantify the steric effects, their
fs = rus (24)
correlations with the indices of processes in which strain plays
dEs imperative role were also explored. Taking the experimental
where the steric potential us is given by us . The steric
dr electrophilicity and nucleophilicity scales and stability of
force shows explicitly the space distribution of the strains in the hostguest complexes as benchmark sets, we showed that from

This journal is the Owner Societies 2016 Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17927

Paper
the viewpoint of steric effects, the information theoretic func-
tionals and quantities derived from their definitions such as
steric force correlate remarkably with the experimental scales of
both electrophilicity and nucleophilicity and stability of com-
plexes. Putting these findings together, it can be concluded that
apart from Fisher information, other quantities from informa-
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
tion theory and especially Shannon entropy can be considered as
quantitative measures of steric effects, wherein further evidences
of the applicability of information theory measures for the
interpretation of various phenomena are highlighted.
Acknowledgements
The authors are grateful to Shiraz University for providing the
needed computing facilities for the present study.
References
1 R. W. Taft Jr., in Steric Eect in Organic Chemistry, ed.
M. S. Newman, Wiley, New York, 1956.
2 R. Homann, THEOCHEM, 1998, 1, 424.
3 P. W. Ayers, J. Chem. Phys., 2000, 113, 10886.
4 S. Shahbazian and M. Zahedi, Found. Chem., 2006, 8, 37.
5 R. G. Parr, P. W. Ayers and R. F. Nalewajski, J. Phys. Chem. A,
2005, 109, 3957.
6 P. W. Ayers, Faraday Discuss., 2007, 135, 161.
7 G. Frenking and A. Krapp, J. Comput. Chem., 2007, 28, 15.
8 P. Hohenberg and W. Kohn, Phys. Rev., 1964, 136, B864.
9 W. Kohn and L. J. Sham, Phys. Rev., 1965, 140, A1133.
10 R. G. Parr and W. Yang, Density Functional Theory of Atoms
and Molecules, Oxford, New York, 1989.
11 S. B. Liu, J. Chem. Phys., 2007, 126, 244103.
12 C. F. von Weizsacker, Z. Phys., 1935, 96, 431.
13 V. F. Weisskopf, Science, 1975, 187, 605.
14 S. B. Liu and N. Govind, J. Phys. Chem. A, 2008, 112, 6690.
15 S. B. Liu, N. Govind and L. G. Pedersen, J. Chem. Phys., 2008,
129, 094104.
16 I. Ugur, F. D. Vleeschouwer, N. Tuzun, V. Aviyente, P. Geerlings,
S. B. Liu, P. W. Ayers and F. De Proft, J. Phys. Chem. A, 2009,
113, 8704.
17 M. Torrent-Sucarrat, S. B. Liu and F. De Proft, J. Phys.
Chem. A, 2009, 113, 3698.
18 S. B. Liu, H. Hao and L. G. Pedersen, J. Phys. Chem. A, 2010,
114, 5913.
19 V. G. Tsirelson, A. I. Stash and S. B. Liu, J. Chem. Phys., 2010,
133, 114110.
20 D. H. Ess, S. B. Liu and F. De Proft, J. Phys. Chem. A, 2010,
114, 12952.
21 R. O. Esquivel, S. B. Liu, J. C. Angulo, J. S. Dehesa, J. Antoln
and M. Molina- Espritu, J. Phys. Chem. A, 2011, 115, 4406.
22 Y. Huang, A.-G. Zhong, Q. Yang and S. B. Liu, J. Chem. Phys.,
2011, 134, 084103.
23 M. Alipour and A. Mohajeri, Mol. Phys., 2012, 110, 2895.
24 L. Rincon and R. Almeida, J. Phys. Chem. A, 2012, 116, 7523.
25 M. Alipour and A. Mohajeri, J. Phys. Org. Chem., 2012, 25, 797.
17928 | Phys. Chem. Chem. Phys., 2016, 18, 17917--17929
View Article Online
PCCP
26 S. B. Liu, J. Phys. Chem. A, 2013, 117, 962.
27 D. Fang, J.-P. Piquemal, S. B. Liu and G. A. Cisneros, Theor.
Chem. Acc., 2014, 133, 1484.
28 C. Rong, T. Lu and S. B. Liu, J. Chem. Phys., 2014, 140, 024109.
29 M. Alipour, Chem. Phys., 2014, 434, 11.
30 W. Wu, Z. Wu, C. Rong, T. Lu, Y. Huang and S. B. Liu,
J. Phys. Chem. A, 2015, 119, 8216.
31 S. B. Liu and C. K. Schauer, J. Chem. Phys., 2015, 142, 054107.
32 S. B. Liu, J. Phys. Chem. A, 2015, 119, 3107.
33 Z. Wu, C. Rong, T. Lu, P. W. Ayers and S. B. Liu, Phys. Chem.
Chem. Phys., 2015, 17, 27052.
34 R. F. Nalewajski, Information Theory of Molecular System,
Elsevier, Amsterdam, The Netherlands, 2006.
35 R. A. Fisher, Proc. Cambridge Philols. Soc., 1925, 22, 700.
36 C. E. Shannon, Bell Syst. Tech. J., 1948, 27, 379.
37 O. Onicescu, C. R. Acad. Sci. Paris A, 1966, 263, 25.
38 S. K. Ghosh, M. Berkowitz and R. G. Parr, Proc. Natl. Acad.
Sci. U. S. A., 1984, 81, 8028.
39 R. F. Nalewajski and R. G. Parr, J. Phys. Chem. A, 2001, 105, 7391.
40 A. Borgoo, M. Godefroid, K. D. Sen, F. De Proft and
P. Geerlings, Chem. Phys. Lett., 2004, 399, 363.
41 R. G. Parr, P. W. Ayers and R. F. Nalewajski, J. Phys. Chem. A,
2005, 109, 3957.
42 K. D. Sen and J. Katriel, J. Chem. Phys., 2006, 125, 074117.
43 P. W. Ayers, Theor. Chem. Acc., 2006, 115, 370.
44 K. D. Sen, J. Antoln and J. C. Angulo, Phys. Rev. A: At., Mol.,
Opt. Phys., 2007, 76, 032502.
45 A. Nagy, Chem. Phys. Lett., 2007, 449, 212.
46 A. Borgoo, P. Geerlings and K. D. Sen, Phys. Lett. A, 2008,
372, 5106.
47 A. Nagy and S. B. Liu, Phys. Lett. A, 2008, 372, 1654.
48 R. F. Nalewajski, Int. J. Quantum Chem., 2008, 108, 2230.
49 J. C. Angulo and J. Antoln, J. Chem. Phys., 2008, 128, 164109.
50 V. G. Tsirelson and A. Nagy, J. Phys. Chem. A, 2009, 113, 9022.
51 A. Nagy and E. Romera, Chem. Phys. Lett., 2010, 490, 242.
52 P. Geerlings and A. Borgoo, Phys. Chem. Chem. Phys., 2011,
13, 911.
53 M. Alipour, Mol. Phys., 2013, 111, 3246.
54 M. Alipour, Chem. Phys. Lett., 2015, 635, 210.
55 A. Nagy, Int. J. Quantum Chem., 2015, 115, 1392.
56 R. O. Esquivel, M. Molina-Espritu, S. Lopez-Rosa, C. Soriano-
Correa, C. Barrientos-Salcedo, M. Kohout and J. S. Dehesa,
ChemPhysChem, 2015, 16, 2571.
57 L. Delle Site, Int. J. Quantum Chem., 2015, 115, 1396.
58 B. R. Frieden, Physics from Fisher Information, Cambridge
University Press, Cambridge, 1998.
59 S. B. Liu, J. Chem. Phys., 2007, 126, 191107.
60 R. G. Parr and L. J. Bartolotti, J. Phys. Chem., 1983, 87, 2810.
61 P. W. Ayers, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 1959.
62 F. De Proft, P. W. Ayers, K. D. Sen and P. Geerlings, J. Chem.
Phys., 2004, 120, 9969.
63 P. W. Ayers, F. De Proft and P. Geerlings, Phys. Rev. A: At.,
Mol., Opt. Phys., 2007, 75, 012508.
64 P. W. Ayers and A. Cedillo, in Chemical reactivity theory: a
density functional view, ed. P. K. Chattaraj, Taylor and
Francis, Boca Raton, 2009.
This journal is the Owner Societies 2016

PCCP
65 C. Rong, T. Lu, P. W. Ayers, P. K. Chattaraj and S. B. Liu,
Phys. Chem. Chem. Phys., 2015, 17, 4977.
66 H. J. Bohorquez, Phys. Chem. Chem. Phys., 2015, 17,
32053.
67 M. Alipour and A. Mohajeri, Mol. Phys., 2012, 110, 403.
68 S. B. Liu, Acta Phys.-Chim. Sin., 2016, 32, 98.
Published on 20 June 2016. Downloaded by UNIV FEDERAL DO RIO DE JANEIRO on 16/08/2016 18:52:40.
69 R. W. Taft, Jr., J. Am. Chem. Soc., 1952, 74, 2729.
70 J. March, Advanced Organic Chemistry, John Wiley & Sons,
New York, 1992.
71 K. C. Hancock, A. E. Meyers and B. J. Yager, J. Am. Chem.
Soc., 1961, 83, 4211.
72 V. I. Talvik and A. V. Palm, Org. React., 1971, 8, 445.
73 H. Mayr and M. Patz, Angew. Chem., Int. Ed. Engl., 1994,
33, 938.
74 H. Mayr, T. Bug, M. F. Gotta, N. Hering, B. Irrgang, B. Janker,
B. Kempf, R. Loos, A. R. Ofial and G. Remennikov, J. Am.
Chem. Soc., 2001, 123, 9500.
75 R. Lucius, R. Loos and H. Mayr, Angew. Chem., Int. Ed., 2002,
41, 91.
76 H. Mayr, B. Kempf and A. R. Ofial, Acc. Chem. Res., 2003,
36, 66.
77 J. W. Steed and J. L. Atwood, Supramolecular Chemistry,
Wiley-VCH, Weinheim, Germany, 2009.
This journal is the Owner Societies 2016
View Article Online
Paper
78 P. Cragg, A Practical Guide to Supramolecular Chemistry,
John Wiley & Sons, Chichester, U.K., 2005.
79 S. M. Bachrach, J. Phys. Chem. A, 2013, 117, 8484.
80 S. M. Bachrach and Z. O. M. Nickle, J. Phys. Chem. A, 2015,
119, 10613.
81 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
82 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785.
83 B. Pinter, T. Fievez, F. M. Bickelhaupt, P. Geerlings and
F. De Proft, Phys. Chem. Chem. Phys., 2012, 14, 9846.
84 M. J. Frisch, G. W. Trucks and H. B. Schlegel et al., Gaussian
09, revision A.02, Gaussian, Inc., Wallingford, CT, 2009.
85 T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580.
86 S. B. Liu, C. Rong and T. Lu, J. Phys. Chem. A, 2014, 118, 3698.
87 S. Kullback, Information Theory and Statistics, Dover, Mineola,
NY, 1997.
88 J. C. Barnes, M. Jurcek, N. L. Strutt, M. Frasconi, S. Sampath,
M. A. Giesener, P. L. McGrier, C. J. Bruns, C. L. Stern,
A. A. Sarjeant and J. F. Stoddart, J. Am. Chem. Soc., 2012, 135, 183.
89 M. Juricek, J. C. Barnes, E. J. Dale, W.-G. Liu, N. L. Strutt,
C. J. Bruns, N. A. Vermeulen, K. Ghooray, A. A. Sarjeant,
C. L. Stern, Y. Y. Botros, W. A. Goddard and J. F. Stoddart,
J. Am. Chem. Soc., 2013, 135, 12736.
Phys. Chem. Chem. Phys., 2016, 18, 17917--17929 | 17929